From USDC settlement to agent-payment evidence

A transaction-level verification study on Base and Solana Public blockchains expose settlement, but claims about agent commerce require additional evidence. This study applies an evidence ladder, reanalysable from preserved raw responses, to native-USDC activity associated with publicly documented PayAI infrastructure on Base and Solana. The frozen convenience sample comprises 25 Base transactions and 100 Solana address-reference signatures acquired on 2 November 2026. Receipt, calldata and token-event checks against provider-reported data verified 25 Base transfers totaling 3.302048 USDC. Solana instruction and historical token-balance reconciliation verified 99 sponsored transfers totaling 1.189720 USDC; the remaining signature was an incoming SOL transfer with a different fee payer. Ten Base cases showed explorer/RPC source agreement (the explorer also supplied candidates); ten Solana cases agreed across two RPC providers that may share upstream data. Unpaid HTTP 402 probes to Echo Merchant endpoints, documented as a refunded test flow, returned native-USDC requirements on both chains, not evidence of completed commerce. Neither sample supplied a transaction-to-request-to-delivery link or verified agent identity. The contribution separates infrastructure-associated settlement, protocol compatibility and evidenced economic activity, with raw responses, deterministic analysis and a coverage record. This is a small measurement audit, not an estimate of market size, adoption or relative chain activity. Independent working paper. Not peer reviewed. This deposit contains the manuscript only; supporting study artifacts are retained privately and may be requested subject to source-use conditions and a separate disclosure review. No application source code or private product documentation is included. The manuscript discloses substantial AI-tool assistance and distinguishes verified USDC settlement from unverified agent-commerce claims.